Representatives from 34 states gathered here last week to discuss options for rating schools based, in part, on the gains that individual students make from year to year.
The meeting, sponsored by the Council of Chief State School Officers’ state-services and technical-assistance division, reflects the mounting enthusiasm for growth or “value added” models that judge schools on how much they contribute over time to students’ learning. (“‘Value Added’ Models Gain in Popularity,” Nov. 17, 2004.)
During the Nov. 15-16 event, participants heard from officials in states that longitudinally track the progress of individual students on state tests as part of their accountability systems, including Florida, North Carolina, and Tennessee.
They also heard from states, such as California, that use a growth model that does not involve tracking individual students from year to year. California’s Academic Performance Index, instead, measures a school’s growth by how well it is moving toward a performance target, according to the results from a composite of tests across grades and subjects.
The state is exploring the “appropriateness and feasibility” of tracking individual students’ achievement growth over time as part of its accountability system. The state department of education is expected to make recommendations to the legislature by next July.
Still, the meeting raised a number of central issues that states must address before they reach that point, including the quality of state data systems, which growth or value-added models are most appropriate for which purposes, and how to communicate results to the public.
No ‘Quick Fix’
“None of these models is a quick fix,” said Mary Yakimowski, the director of assessments for the state chiefs’ council. “There are pluses and minuses.”
Two big issues raised at the meeting were how to decide how much growth is enough, and how to ensure that state tests can accurately measure students’ development across grades.
In particular, attendees wondered how to craft systems that reward schools for growth, while continuing to hold them accountable for bringing students to the “proficient” level on state tests, as the federal No Child Left Behind Act requires.
Otherwise, low-performing students could conceivably make a year’s worth of growth each year and still never reach a proficient level, based on state achievement standards, before they graduated.
State testing directors also discussed how to align state academic-content standards and tests across grades, in order to measure students’ progress from grade to grade. That’s a particular challenge if the content being measured changes from one grade level to the next—for example, from computation to algebra.
Peter Goldschmidt, an economist at the National Center for Research on Evaluation, Standards, and Student Testing at the University of California, Los Angeles, pointed out that value-added models could potentially provide states with a wealth of data about their schools, not just a single number. Such models could be used, he said, to examine whether some schools produce higher growth rates for poor or minority students than other schools do. States or districts could then examine what the more effective schools are doing differently.
The council plans to continue working with states on various growth or value-added models, perhaps by sharing information across states or by publishing a paper that describes the strengths and weaknesses of different approaches.